Geospatial Optimization Analytics on Hotpot Site Selection in Beijing Based on Deep Reinforcement Learning

The accelerating process of urbanization and the subsequent enhancement of living standards among urban residents have resulted in a persistent increase in demand for hotpot restaurants. Conventional location selection methodologies have demonstrated deficiencies in addressing the intricate market demands, dynamic changes, and multi-objective optimization, resulting in challenges such as imbalanced resource allocation, inadequate coverage, and suboptimal computational efficiency. To address these challenges, this study proposes a novel approach that combines the Maximum Covering Location Problem, genetic algorithms, Gurobi optimization, and deep reinforcement learning to optimize site selection for hotpot restaurants within the Sixth Ring Road of Beijing. The findings suggest that deep reinforcement learning surpasses genetic algorithms in terms of computational efficiency and coverage performance, effectively achieving resource balance and spatial coverage while circumventing prevalent issues such as over-concentration and resource waste. Notwithstanding its limitations, the proposed method offers novel theoretical insights and practical references for the optimization of hotpot restaurant location selection. Future research may further incorporate additional influencing factors, refine weight calculation methods, and enhance the accuracy and practicality of location optimization.